Hammill: Outgassing and TiSP 2
Outgassing of Stainless Steel Vacuum Chambers and
The Vacuum Pumping Performance Evaluation of a Titanium
Sublimation Pump
Christian Hammill a)
Department of Physics and Astronomy, Wayne State University, Detroit, Michigan, 48201
Reducing material outgassing is essential to achieve an extremely high vacuum (XHV). Previous studies have shown extremely low outgassing rates from stainless steels (≈2x10-14 Torr•L•s-1•cm-2) after they have been baked at 400°C in either a vacuum or air. This first project re-examines stainless steel outgassing rates on a pretreated (baked) stainless steel chamber that has been stored properly under N2 for ~8 months. The outgassing rate of the sample chamber is determined mainly via the rate of rise method. The results of the rate of rise method are cross checked via the throughput method. Results have indicated that the stainless steel chamber maintained its extremely low outgassing rate after long term storage.
A Titanium Sublimation Pump (TiSP) will be used to handle very large gas loads at the entrance of the beam dump for the Cornell Prototype ERL Photo-cathode injector. The second part of this study is to evaluate the vacuum pumping performance of the TiSP, specifically the pumping speed and capacity. Our results have shown the TiSP to have a maximum pumping speed of >1,000 L/s and a pumping capacity ≤25 Torr•L for H2 gas.
Hammill: Outgassing and TiSP 2
I. INTRODUCTION
This REU project consists of two parts: the study of the outgassing of a stainless steel chamber that was previously treated, tested, and stored in N2, and the pumping performance evaluation of a titanium sublimation pump (TiSP). Both projects are closely related to the R & D efforts in construction of the Cornell Proto-type Photo-cathode Injector1, whose purpose is to shape and study the properties of the electron beam, as it must fit certain characteristics for the Energy Recovery Linac (ERL).
In certain places of the ERL, like the photo-cathode, an XHV must be maintained (≤10-12 Torr). Reducing material (mainly SST) outgassing is essential to achieve an XHV. Previously, outgassing properties of stainless steel vacuum chambers have been studied. It has been found that treating a stainless steel system at 400°C can dramatically reduce H2 outgassing2.
______
a) Electronic mail:
The goal of the first part of the project is to find whether or not this extremely low outgassing rate lasts after the treated chamber has been properly stored in N2 for a significant period of time (nearly 8 months in this case). This is done by examining the current outgassing rate of the sample chamber and comparing it to its outgassing rate prior to storage. This is a replication of a practical scenario in which the vacuum chamber is vented for repairs and/or updates, and then parts used on the chamber are properly stored in N2. The outgassing properties of the sample chamber were then reexamined.
The second part of this REU project deals with the vacuum pumping issue in the Cornell ERL project. After measuring the properties of the electron beam generated in the photo-cathode electron gun, the beam must be safely terminated at the aluminum beam dump (see Fig. 1). Very large gas loads (predominately H2 gas) are generated at the beam dump due to electron induced desorption at the aluminum surface in the dump.
A large TiSP will be used together with two large ion pumps to control the H2 gas load. The TiSP chamber was chosen as it was already within the possession of the lab. So, utilizing it will save both time and money.
In this part of this project, we will evaluate the pumping performance of the TiSP, which is known to be very effective in pumping H2 gas3. The results of these tests will help to determine whether this TiSP is suitable for the application.
Fig. 1- A display depicting the Cornell Prototype ERL Photo-cathode injector beam line after the superconducting cavities. Marked in circles are the areas of importance to this paper, (i.e., the TiSP chamber and the Beam Dump chamber).
II. METHODOLOGY
A. Stainless steel outgassing measurements
As stated earlier, the mission of this part of the project is to re-measure the H2 outgassing rate, , of a stainless steel (type 314L in this case) chamber. This chamber was previously treated with a 400°C vacuum bakeout and was then stored in N2 for ~8 months.
The outgassing measurements are carried out in a setup shown in Fig. 2. The vacuum system as a whole is divided into two parts: the testing chamber and the sensor chamber. These two components are connected through an all-metal angle valve just outside the oven. It is necessary to divide the whole vacuum system into two parts, because we need independent controls of the sensor chamber and the testing chamber. Using this setup, the stainless steel outgassing rates are measured via two methods: the rate of rise method and the throughput method.
Fig. 2- A sketch depicting the outgassing vacuum system with chamber, SRG, and various vacuum components.
The testing chamber consists of the oven and all the contents contained therein. The pretreated sample (stainless steel) chamber is placed in the oven. During the rate of rise method, the temperature in the oven is controlled to within 0.1°C of a setting temperature of 25°C by aid of a combination of heating tape and water cooling coils. The heating tape and the water cooling coils are controlled outside the vacuum system via a PID controller, and the pressure in the chamber is monitored by a spinning rotor gauge, SRG.
The sample chamber and the SRG sensor head are placed inside the oven, which is designed to bake the sample chamber uniformly up to 450°C. The oven is capable of baking at this temperature by way of a high power heating gun and air blower.
The second part of the setup, the sensor chamber, consists of a cold cathode gauge, (CCG), a residual gas analyzer, (RGA), and a small ion pump. A turbo molecular pump is then connected to the sensor chamber. The sensor chamber is used for throughput measurements, determination of gas composition, and for pumping down.
In this study, the testing chamber is baked at various temperatures starting at 150°C and up to 250°C, while the sensor chamber is consistently baked at 250°C during each trial run by being wrapped in heat tape. This temperature will eliminate water, hydrocarbons, and other carbon compounds like CO and will achieve a base pressure in the low 10-10 Torr range.
The pressure in the sensor chamber is monitored by the CCG. A RGA determines the makeup of the gas in the system. A turbo molecular pump is used for an initial pump-down and to handle large gas loads during bakeouts. The small ion pump is used to maintain the ultra high vacuum and for the throughput method.
The sensor chamber is made as small as possible so that its contribution to the outgassing load will be minimized.
All of the measurements, regardless of the method of choice, involve a similar procedure:
Setup and leak check: First, all of the vacuum flanges are connected properly and tightened. The entire system is then pumped down via the turbo molecular pump. Once pumped down, the system is checked for leaks with a RGA using helium as a trace gas.
Bake-out: Whenever a stainless steel vacuum system is exposed to air, water molecules adsorb onto the chamber surface. The adsorbed water must be eliminated via bakeout at temperatures ≥120°C.
The testing chamber is baked in the oven using heat gun and a hot air blower. The temperatures of both heating systems are controlled using programmable PID controllers. A typical bakeout cycle includes a smooth ramping up to a desired temperature, TB, holding at TB for a requested duration, and a smooth ramping down. Once the whole vacuum system has baked for a period of 48 to 72 hours, the system is left to cool and leak checked again to see if the bakeout caused any of the components in the apparatus to leak.
Measurement: When utilizing the rate of rise method, the testing chamber is isolated from the sensor chamber, and the SRG is used to measure the pressure in the testing chamber. When utilizing the throughput method, the valve to the testing chamber opens to the sensor chamber, and the pressure is measured with a CCG. The makeup of the gas is then determined using a RGA.
Both the CCG and SRG controllers are interfaced to a PC with a 12-bit data acquisition card (National Instruments© DAQ-1200). The recorded analog output voltages, (in volts) of both gauges are later converted into the gauge pressure (in torr) using the following:
PSRG=1x10-6• (1)
PCCG= (2)
Repeat: This series of steps is repeated, only each time, the baking temperature is raised. The first time this cycle of steps was ran, the baking was 150°C. It was then was raised to 200°C in the next cycle and then to 250°C in the last cycle. These repetitions, which only vary by the baked temperature, help to ensure accuracy of the data.
A1. Rate of Rise Method
In the rate of rise method, the (stainless steel) sample chamber is completely closed off to allow gas accumulation. Without any pumping in the testing chamber, and assuming a constant outgassing rate, , the accumulation of the outgassing should result in a linear pressure increase over time, that is, a constant rate of rise in pressure, . The outgassing rate, , can be calculated by:
•V/As (3)
Where V is the volume of the chamber (28 L) and As is the inner surface area of the chamber (7500 cm2).
The SRG4 is chosen for the rate of rise method primarily because the gauge does not alter (or disturb) the vacuum to be measured. This is distinctly different from the other types of gauges (such as the cold cathode gauge, CCG), which may introduce a small amount of pumping and/or may also cause a small outgas.
Fig. 3- The arrangement of a typical Spinning Rotor Gauge apparatus.
The SRG consists of ball (rotor) in a thimble-like chamber (gimble) (see Fig. 3) in which it is suspended via a magnetic field. The rotor then begins to spin and eventually “settles” at a constant spinning frequency. Molecules in the chamber collide with the rotor, inducing a molecular drag. The amount of molecular drag on the rotor is directly proportional to the pressure inside the chamber.
It is very important that the magnetic levitation force align with gravity to ensure a stable rotor rotation. This is achieved by using a small level to adjust the orientation of the SRG head.
Other factors can affect the spinning frequency as well. For instance, temperature can cause the ball to either expand or shrink. This directly affects the inertia of the ball and therefore indirectly affects the spinning frequency of the ball. So, in all of the experiments, cooling coils and a heater (see Fig 2) were used to maintain the temperature inside the oven within ∆T≈ 0.1°C of the set temperature of 25°C.
Meaningful outgassing data was only obtained at night and/or over the weekend, because the SRG is highly sensitive to vibrations. Walking around the chamber (by lab personnel) or even lightly touching the system can cause the SRG to produce high amounts of “noise”.
While these sources of error have been minimized, the SRG can only reliably measure pressure >10-7 Torr. This is a limit set by the residual magnetic drag. In our setup, the pressure is further limited to >10-6 Torr, due to lack of vibrational isolation to minimize ground motion. The scatter in the data is mostly caused by the residual ground vibration.
A2. Throughput method
As a cross check to the rate of rise method, the throughput method is employed. This method uses a change in pressure, ∆P (as explained later) and the pumping speed of the ion pump, Sip, such that:
= ∆P•Sip/As (4)
Using a CCG and having the chamber sealed, we measure a baseline pressure in the sensor chamber. Then the testing chamber is opened to the sensor chamber to pump out the accumulated gas, and the pressure is allowed to settle to equilibrium. The absolute difference between the settled pressure and the baseline pressure is then recorded as our ∆P. Upon opening the valve between the testing chamber and the sensor chamber, the accumulated gas (predominately H2; see Fig. 4b)) in the testing chamber is quickly pumped away by the ion pump, as shown in Fig. 4a). This pump down curve, i.e. pressure vs. time, can be fitted by the following6:
f(t)= Po + P1e∆t/τ1 + P2e∆t/τ2 (5)
Where, Po, P1, P2, τ1, and τ2 are fitting parameters (See Appendix).
a)
b)
Fig. 4- a) Typical pump down curve as measured by the CCG
b) The mass spectrum of gas at the beginning of the pump down (marked as point A in a)). This shows the dominance of H2 gas.
The pumping speed of the ion pump, Sip, can be experimentally determined by analyzing the pumping down transient. One can relate the Sip to the fitted τ1 as:
Sip=V/τ1 (6)
It is worth noting that the pumping speed determined in this way is independent of gauge calibration.
B. Determination of TiSP vacuum properties
In previous sections, it was mentioned that the purpose of the TiSP is to capture a large gas load generated in the aluminum beam dump. In Fig. 5, one can see when the beam is dumped, it causes in increase in pressure of up to ~2x10-5 Torr. In order to control this pressure increase and to control the gas load, a TiSP along with 2 ion pumps were placed in the A5 section.
Fig. 5- The calculated5 pressure distribution along the ERL Beamline, given the conditions that the A5 pumping speed=2000L/s and electron induced desorption yield, YEID=0.1. See also Fig 1.